The rise of AI-assisted lawmaking

Artificial intelligence is beginning to influence not only how governments enforce and interpret laws, but also how they write them. What was once largely the domain of legislators, policy analysts, and legal experts is becoming a new frontier for AI deployment as governments experiment with tools capable of drafting legislative language, analyzing policy proposals, identifying inconsistencies across legal codes, and forecasting the effects of regulatory decisions.

The trend reflects a broader expansion of AI throughout the public sector. The Organisation for Economic Co-operation and Development's (OECD) September 2025 report Governing with Artificial Intelligence identified nearly 200 examples of AI use across government functions worldwide, ranging from regulatory design and policy evaluation to administrative decision-making. While most of these systems remain advisory rather than autonomous, their growing presence raises an important question. If AI can help governments administer laws more efficiently, should it also help create them?

The answer carries implications far beyond technological innovation. Laws determine how societies allocate resources, define rights, regulate markets, and exercise public authority. As governments confront increasingly complex policy challenges, many are exploring whether AI can augment legislative capacity. At the same time, the technology raises difficult questions about transparency, bias, and democratic oversight that policymakers are only beginning to address.

Expanding legislative capacity

For many governments, the appeal of AI-assisted lawmaking lies not simply in speed but in scale and complexity.

According to a 2025 analysis published by Lawfare, a prominent publication focused on law and public policy, large language models possess two characteristics that could make them particularly useful in legislative drafting. First, they offer expansive scope, allowing lawmakers to work across multiple specialized policy areas simultaneously without requiring large teams of subject-matter experts. Second, they can process levels of complexity that exceed human capacity, reviewing vast quantities of legislation, regulatory guidance, legal opinions, and policy documents while maintaining consistency across interconnected legal frameworks.

Those capabilities address a growing challenge facing modern legislatures. New laws must often account for existing statutes, judicial decisions, and overlapping responsibilities among government agencies. Reviewing those materials manually can require months of legal and policy analysis. AI systems offer the possibility of accelerating that process by identifying redundancies, omissions, and potential conflicts before legislation reaches a vote.

That potential is already attracting attention from policymakers. The OECD report notes that governments are beginning to explore AI applications in areas such as legislative drafting and regulatory analysis. In most cases, the technology is not intended to replace lawmakers, but to assist civil servants and legal experts by generating draft language, reviewing regulations, and supporting policy assessments.

Experiments are already underway in several legislative institutions. In Brazil, parliamentary bodies have explored AI-assisted systems that help lawmakers analyze legislation, compare legal texts, and navigate large volumes of parliamentary documentation. Similar efforts have emerged in the United Kingdom, where parliamentary institutions have tested generative AI tools to support research and document analysis.

These early initiatives suggest that AI's role may extend well beyond drafting assistance. A 2025 study published in the journal The Theory and Practice of Legislation argues that the technology could become integrated throughout the legislative process, assisting policymakers not only with writing legislation but also with researching issues, evaluating alternatives, and assessing the likely consequences of policy decisions.

Accountability in the age of algorithmic governance

The growing use of AI in legislative processes has prompted concerns about who bears responsibility when automated systems influence public policy.

Unlike traditional legal research tools, generative AI systems can produce novel text and recommendations that may contain factual inaccuracies, embedded biases, or flawed assumptions. If such outputs contribute to legislation, determining accountability becomes more complicated.

The Inter-Parliamentary Union (IPU), a global organization representing national parliaments, identified transparency, accountability, and public trust among the principal governance challenges associated with parliamentary AI systems in its 2024 guidelines The Role of AI in Parliaments. Democratic legitimacy depends in part on citizens understanding how laws are developed and who is responsible for their contents. When AI-generated analyses or draft provisions become part of that process, maintaining that transparency may become more difficult.

Questions also arise regarding the data used to train these systems. Bias compounds these concerns. The OECD notes that AI models trained on historical data may reproduce or amplify existing biases embedded within those datasets. In legislative contexts, that could mean reinforcing longstanding inequalities or policy assumptions rather than challenging them.

Concerns extend beyond bias alone. In its 2025 analysis, Lawfare argues that if sophisticated AI systems become central to legislative drafting, influence over the design, training, and operation of those systems could itself become a source of political power. Critics warn that AI tools optimized to serve particular interests could entrench powerful constituencies more effectively than traditional lobbying mechanisms.

For these reasons, many researchers and policymakers emphasize that AI should remain a decision-support tool rather than a decision-maker. Human oversight remains central to ensuring that legislative choices reflect democratic priorities rather than algorithmic outputs.

The future of lawmaking

For now, fully automated lawmaking remains firmly in the realm of speculation. Legislators continue to write, debate, amend, and approve laws, while AI serves primarily as an analytical assistant.

Nevertheless, governments around the world are actively testing how these technologies can support increasingly complex legislative and regulatory responsibilities. The question is no longer whether AI will enter the lawmaking process. In many jurisdictions, it already has.

Whether that development ultimately strengthens governance will depend less on the technology itself than on the safeguards surrounding its use. Transparency requirements, human oversight mechanisms, and independent review processes may prove just as important as the underlying algorithms.